Automated operational modal analysis based on long‐term records: A case study of Milad Tower structural health monitoring

Structural modal identification is time-consuming for large-scale structures with continuous data recordings and long-term data collections. This paper proposes an automated three-step procedure to simplify the identification and monitoring of modal parameters. In the first step, the recorded data are divided into small segments and evaluated in energy and frequency contents to select high-quality segments. In the second step, the high-quality segments are utilized in a simplified stochastic subspace identification algorithm to determine the modal parameters. Finally, the modal parameters' statistical features and uncertainty bounds are extracted for structural health monitoring. The procedure is implemented on Milad Tower in Tehran as a real-world case study. The modal parameters are identified based on the data continuously recorded for 3 years, and their variations are publicly reported for the first time. The effectiveness of the simplified subspace algorithm is also examined versus the frequently used alternatives. Several identification scenarios are planned based on different sensor arrangements on the tower. Carrying out a statistical study reveals a slight change in Milad Tower's natural frequencies. The findings confirm the capability of the proposed procedure for large-scale structures and digital records for long-term structural health monitoring.